Investigation of the cancer testis antigen lactate dehydrogenase C as a CD8 T cell target

(1)

Investigation of the Cancer Testis Antigen Lactate Dehydrogenase C as a CD8 T Cell Target

by

David S Neilson

Bachelor of Science, University of British Columbia, 2012

A Thesis Submitted in Partial Fulfillment of the Requirements for the Degree of

MASTER OF SCIENCE

in the Department of Biochemistry and Microbiology

©David S Neilson, 2016 University of Victoria

(2)

ii Investigation of the Cancer Testis Antigen Lactate Dehydrogenase C as a CD8 T Cell Target

by

David S Neilson

Bachelor of Science, University of British Columbia, 2012

Supervisory Committee Dr. Julian J. Lum, Supervisor

Department of Biochemistry and Microbiology

Dr. John Webb, Departmental Member Department of Biochemistry and Microbiology

Dr. Francis Nano, Departmental Member Department of Biochemistry and Microbiology

Dr. Ben Koop, Outside Member Department of Biology

(3)

iii

Abstract

The infrequency of known T cell targets in high grade serous ovarian carcinoma (HSGC) is a substantial barrier to the development of targeted immunotherapies. Due to their infrequency, antigen discovery is a crucial component of immunotherapeutic design. In our cohort of HGSC cases, the cancer-testis (CT) antigen lactate dehydrogenase C (LDHC) is expressed in 76% of tumours (22/29). As LDHC presents with tumour specificity in women, I hypothesize that LDHC is an immunogenic target in HGSC patients, and that LDHC-specific T cells can be activated and expanded for therapeutic purposes. As such, I sought to examine whether endogenous LDHC-specific T cells were present in the ascites of HGSC patients. A standard Rapid Expansion Protocol was used to expand CD8 T cell cultures from patient ascites. These cultures were screened for reactivity to a peptide library encompassing all possible epitopes of the LDHC protein by interferon-γ ELISpot. With this approach, T cell clones from one of five patients were identified that were reactive to minimal peptides contained within LDHC. In this patient, the antigenic LDHC peptide differentiated from LDHA by a single amino acid at its C-terminus

(YTSWAIGLSVM versus YTSWAIGLSVA). In recognition assays, tumour cell lines expressing endogenous LDHC, autologous ascites, or autologous B cells transfected with LDHC were unable to elicit T cell responses. Although this study suggests that LDHC is not immunogenic, continued screening of LDHC and other CT proteins will likely provide additional immunotherapeutic targets.

(4)

iv

List of Tables

Table 1. LDHC peptide library. ... 38 Table 2. LDHC peptide 62 minimal peptide library and its LDHA complement. ... 39 Table 3. Compatibility of Patient 2, OVCAR5, and OVCAR8 HLA haplotypes. ... 76

(8)

viii

List of Figures

Figure 1. Antigen presentation by MHCI to CD8 T cells. ... 10

Figure 2. Alignment of LDHC to LDHA and LDHB.. ... 27

Figure 3. Patient T cell responses to pooled LDHC library peptides as evaluated by IFNγ ELISpot ... 48

Figure 4. Post-expansion IFNγ ELISpots of first-pass reactive T cells ... 50

Figure 5. Fluorescence activated cell sorting of CD8+ 4-1BB+ cells after overnight incubation with cognate peptide ... 53

Figure 6. Secondary isolation of Patient 2 LDHC peptide 62 reactive T cells. ... 56

Figure 7. Recognition of LDHC p62 minimal peptides by D4 T cells as evaluated by IFNγ ELISpot ... 57

Figure 8. Purity of the Patient 2 D4 culture as assessed by TCR-Vß spectratyping ... 60

Figure 9. OVCAR recognition assay by Patient 2 D4 T cells as evaluated by IFNγ ELISpot... 76

Figure 10. Ascites recognition assay by Patient 2 D4 T cells as evaluated by IFNγ ELISpot ... 78

(9)

ix

Acknowledgements

Firstly, I would like to thank Julian Lum for taking me on as a Master’s student in his lab. This event led me to an experience in Victoria which I could never have seen coming. It has been a few years of great opportunity and learning. Julian took a leap of faith by taking me on, and I will always appreciate that. I also must thank the members of my graduate committee: Francis Nano, Ben Koop, and John Webb. It was you in part who ensured that I proceed appropriately through my degree and jump through the correct hoops at the correct times.

To all the students and scientists at the Deeley Research Centre, thank you for countless sessions of input over the span of three years. Our interactions over the years, both in and outside of the lab, have been invaluable in shaping my project and academic person. Thank you for the challenges and support you have offered.

I would like to specifically credit Darin Wick, who was a mentor of sorts to me throughout the first half of my studies. Beyond instructing me in technical aspects of lab science and experimental design, Darin encouraged me to develop a more inquisitive and analytical mind, whilst also challenging my personal growth.

Additionally, Jennifer Kalina contributed a significant amount to my work here at the DRC. In May 2014, Jenn began a coop term working alongside me in the lab and immediately proved how driven and capable she was in the lab. Throughout her time here she has been irreplaceable and has provided me with invaluable input and support.

Additionally, I’d thank my family, who has been endlessly supportive since long before I considered earning a Master’s degree. To my parents, Scott and Valerie Neilson, thank you for being amazing. Thank you for teaching me to live ethically and to care for the people around me. Thank you for encouraging me in my pursuits regardless of what they may have been and always being there if they didn’t work out.

To my brothers, Brad and Ted Neilson, thank you for being brothers. Thank you for all the games and fights and talks we’ve been through, because those experiences have only strengthened our brotherly love. It’s hard to qualify what I’m truly thankful for, but you each have inspired me in unique ways throughout our lives and over the last few years. There will always be new things to learn from you and I look forward to what the future may bring.

To the ultimate community of Victoria, thank you for accepting me with open arms. Being a part of this community has been instrumental in shaping who I am and how I live my life. I’ll also thank my

teammates from the UVic Men’s ultimate team. It has been a pleasure training and competing alongside you. Specifically to my former captains and co-captains at UVic, 22, 50, 34, 6, 73, 45, 16, thank you for tolerating my intense academic demands which occasionally came at a cost to the team.

(10)

x

Dedication

This thesis is dedicated to those who are afflicted and affected by cancers of all origin. May continued research at labs and clinics throughout the world prove to be beneficial to those in need.

(11)

Chapter 1: Introduction

1.1 Prologue

This thesis investigates the antigenicity of lactate dehydrogenase C, in order to elucidate its potential role in a targeted immunotherapy of ovarian cancer. The purpose of this

introductory chapter is to provide context for this thesis and the work it describes. To begin, in the first section I address the frequency and effects of ovarian cancer, as well as the treatments that are most frequently provided to ovarian cancer patients. To better illustrate the interaction of the immune system with cancer, I provide an overview of numerous T cell subsets, which direct the cell mediated adaptive immune response, as well as a short review of T cell maturation and intracellular antigen processing. I then examine a model which depicts the immune system’s interaction with a developing cancer and describe how this interaction engineers a tumour that is able to survive and proliferate despite the presence of a healthy immune system. I go on to discuss immunotherapeutic treatments and their efficacy. Lastly, I discuss a prospective immunologic target: lactate dehydrogenase C. I review its enzymatic activity, its isoforms, and potential reasons for its common expression in cancer.

1.2 Prevalence and Treatment of Ovarian Cancer 1.2.1 The effect of ovarian cancer in Canada

Cancer is an extremely wide reaching and devastating disease. In 2015, over 200 000 Canadians were diagnosed with cancer (Canadian Cancer Society’s Advisory Committee on Cancer Statistics, 2015). Although often described as a single disease, cancer pathology is diverse and

(12)

2 varies among cases. A cancer may arise from virtually any nucleated cell in the body, and each cell type may promote a cancer with a different phenotype. The most accessible classification for malignant tumours is by the local tissue it has arisen from. A mere 1.4% of all cancer cases are classified as ovarian cancer. While this incidence is proportionally quite low, ovarian cancer represents 15 new cases in every 100 000 women per year. In Canada alone, over 2700 women are diagnosed with ovarian cancer each year, with the majority of them presenting with late-stage advanced disease. The mortality rate of ovarian cancer is higher than that of any other gynecological cancer, with an overall five-year survival rate of approximately 40%. In 2015, ovarian cancer led to the death of 1739 Canadian women (Canadian Cancer Society’s Advisory Committee on Cancer Statistics, 2015). Its lethality justifies further research into HGSC therapy.

1.2.2 Subtypes of ovarian cancer

Ovarian cancer is further classified into a set of histotypes, which vary in pathology and prognosis. All ovarian cancers are divided into the broad categories of epithelial and non-epithelial, based on cell of origin. Non-epithelial malignancies may be further divided into germ cell tumours and sex cord stromal tumours (Colombo et al., 2012). The rare non-ovarian

histotypes are commonly diagnosed at early stages, leading to relatively favourable outcomes and relatively positive prognosis (Colombo et al., 2012; Smith et al., 2006). Epithelial ovarian cancer (EOC) constitutes approximately 90% of ovarian cancer cases and consists of five

primary histological subtypes: clear-cell, endometrioid, mucinous, low-grade serous carcinoma, and high-grade serous carcinoma (HGSC) (Sopik, Iqbal, Rosen, & Narod, 2015). The high

(13)

3 mortality rate of ovarian cancer is primarily attributable to HGSC, which at 75% is the most frequently occurring histotype of EOC (George, Garcia, & Slomovitz, 2016). The early stages of HGSC lack observable symptoms, clinically masking the disease and frequently leaving it undetected prior to progression to an advanced stage. When it is detected in its early stages, the five-year survival rate is greater than 90%, while fewer than 20% of advanced stage HGSC patients survive to five years beyond their initial diagnosis (Cuellar-Partida et al., 2016; George et al., 2016; Malpica et al., 2004). Individual histotypes of ovarian cancer must be individually researched so as to provide the best therapy possible.

1.2.3 Treatments of epithelial ovarian cancer

Currently, the standard treatments available for EOC are lacking in efficacy. The front-line standard of care treatment regimen for ovarian cancer is tumour debulking surgery

followed by platinum- and taxane-based chemotherapy (Bookman, 1999). There have been no substantial improvements in this standard treatment in the past 25 years (Bookman, 2016). A high proportion of patients (up to 80%) respond well to front-line treatment; however, disease recurs in the majority of patients (Bookman, 1999; Davis, Tinker, & Friedlander, 2014). Tumours that recur prior to six months post-treatment are deemed ‘platinum-resistant’ (platinum

resistant ovarian cancer; PROC), while those recurring after six months are deemed ‘platinum-sensitive’. Although this qualification is defined at a relatively arbitrary point in time on a continuum, this division has broadly separated patients in a clinically relevant way. Patients with platinum-sensitive tumours are provided additional cycles of platinum based treatment,

(14)

4 which vary in efficacy. Response rates highly correlate with the previous time to recurrence (Bookman, 1999). As few as 30% of patients with platinum-sensitive tumours which have recurred within 1 year respond well to additional chemotherapy cycles, while up to 90% of those which have recurred beyond 1 year may respond well to second-line platinum therapy (Blackledge, Lawton, Redman, & Kelly, 1989; Markman et al., 1991). Despite varied response rates and therapeutic sensitivity, progression free survival is extremely rare – with few exceptions, all recurrent tumours eventually become platinum resistant.

Once a HSGC tumour has been deemed platinum resistant, alternative

chemotherapeutic agents are offered to patients for treatment. While there are numerous options, including paclitaxel, topotecan, and gemcitabine, none offer patients a high probability of tumour regression or increased survival. It is estimated that, at this point, response rates to non-platinum-based chemotherapies have fallen to between 7 and 22% (Thigpen, 2012). Patients may also opt for treatment with tumour targeted agents or enroll in a clinical study – some practitioners believe that some may be better options than non-platinum chemotherapy. Indeed, a select few of these targeted therapies have demonstrated clinical efficacy in HGSC and have been approved by the Food and Drug Administration (FDA) for human use. The agent bevacizumab, an inhibitor of VEGF-angiogenesis, has recently been found to improve response rate, progression free survival, and overall survival compared to standard chemotherapy, in a series of phase III trials (Marchetti et al., 2016). Olaparib, an inhibitor of poly ADP ribose polymerase (PARP), an enzyme involved in DNA repair, has also been approved by the FDA for recurrent HGSC. Studies show that, alone or in combination with paclitaxel and carboplatin,

(15)

5 Olaparib significantly improves progression free survival (Ledermann et al., 2012; Meehan & Chen, 2016; Oza et al., 2015). In addition to the development of targeted cancer therapies, a variety of immunotherapeutic techniques have been explored, some demonstrating efficacy mediated by CD8 T cell activity.

1.3 Immunological T cell Activity

As the primary effector of the cell-mediated immune system, CD8 T cells are responsible for the elimination of infected and malfunctioned cells. All CD8 T cells have unique T cell

receptors (TCR) with defined specificity and affinity for protein fragments, called epitopes, which are held by the binding groove of the major histocompatibility complex class I (MHCI), a highly polymorphic surface molecule expressed by all nucleated cells. Endogenous proteins are routinely digested for epitope display by MHCI. As such, these epitopes may be naturally occurring, derived from DNA mutations, or of viral or bacterial origin. Proteins that generate immunogenic epitopes are deemed antigens. In the periphery, naïve CD8 T cells become

activated upon TCR engagement by an antigen presenting cell (APC), when accompanied by the appropriate costimulation. To achieve activation, the TCR:MHCI interaction is stabilized by CD8 and is accompanied by engagement of the T cell surface receptor CD28 by the APC ligand B7 (Lenschow, Walunas, & Bluestone, 1996). Activation triggers clonal T cell proliferation and the generation of an antigen specific effector population.

Activated CD8 T cells exhibit a targeted killer phenotype and may eliminate an antigen specific cell population. Upon target recognition, perforin and granzymes are released via

(16)

6 exocytosis of cytoplasmic granules (Harty, Tvinnereim, & White, 2000). Perforin molecules form pores within the target cell membrane, allowing cellular infiltration by granzyme B, which initiates apoptosis in the target cell by caspase activation. Alternatively, CD8 T cell effector function is also mediated by Fas/FasL interaction, whereby T cells expressing FasL bind to Fas on the surface of target cells (Harty et al., 2000). Fas engagement promotes the recruitment of the death-induced signaling complex (DISC), subsequently activating the caspase cascade and causing apoptosis. Additionally, CD8 T cells release a number of cytokines upon TCR

engagement. Amongst others, cytokines they release include interferon-γ (IFNγ) and tumour necrosis factor α (TNFα), which have been shown to induce anti-tumour activity (Williams & Bevan, 2007). Indeed, tumour infiltration by CD8 T cells correlates with improved prognosis in various cancer types (Hamanishi et al., 2007; Milne et al., 2009; Sato et al., 2005; Zhang et al., 2003).

1.3.1 Development of TCR specificity during T cell maturation

T cell maturation begins with the generation of TCR variability. This process begins in the bone marrow, where all lymphocytes originate as haematopoietic progenitor cells. Some of these progenitor cells migrate to the cortex of the thymus, where they expand into a population of thymocytes. To generate TCR diversity, early thymocytes undergo random genetic rearrangements at the TCR α and ß loci, which together encode the heterodimeric TCR (Fulton et al., 2015). Due to this genetic rearrangement, each T cell progenitor is committed to one of 1015 possible TCR arrangements (Nikolich-Zugich, Slifka, & Messaoudi, 2004). At this

(17)

7 point, thymocytes also express the surface glycoproteins CD8 and CD4, which stabilize

interaction of the TCR with either MHCI or major histocompatibility complex class II (MHCII), respectively. Thymocytes are then driven toward either a CD8 T cell lineage or a CD4 T cell lineage. Also known as T helper cells, CD4 T cells recognize epitopes in the context of major histocompatibility complex class II (MHCII), which is expressed only by APCs. As TCR generation is random, it must be determined if the TCR can recognize MHCI, MHCII, or neither. Cortical thymic epithelial cells (cTEC) display self-antigen derived peptides in the context of MHCI and MHCII. While thymocytes migrate through the thymic cortex, they bind any MHC:peptide complexes that have a high affinity for their TCR (Klein, Kyewski, Allen, & Hogquist, 2014). Successful binding of self-antigen provides a survival signal, while a lack of signal leads to death by neglect. Thymocytes that receive a survival signal also commit to a CD4 or CD8 single

positive lineage, in accordance with TCR binding to MHCII or MHCI, respectively.

After commitment to a CD4 or CD8 phenotype, thymocytes migrate to the medulla, where tolerance to self-antigen is induced by negative selection. Immunological tolerance of tissue-restricted proteins requires that all potentially autoreactive thymocytes are deleted. To this end, medullary thymic epithelial cells (mTEC) express the transcription factor autoimmune regulator (AIRE), which promotes localized ectopic expression virtually all tissue-restricted transcripts (Klein et al., 2014; Malchow et al., 2016). At this stage, if thymocyte TCRs bind MHC:peptide displayed by mTECs with high affinity, the thymocyte receives an apoptotic signal and is thus deleted (Gallegos & Bevan, 2004; Hinterberger et al., 2010; Oukka,

(18)

8 thymocytes which exhibit an intermediate affinity for MHCII:peptide receive signals to

differentiate into T regulatory (Treg) cells (Apostolou, Sarukhan, Klein, & von Boehmer, 2002; Jordan et al., 2001). Mature T cells then migrate from the thymus into the peripheral blood and lymph where they can initiate their directed effector function.

The efficacy of thymic selection decreases the likelihood of mature T cells recognizing tissue-restricted proteins when they are expressed by tumours. However, despite the vast expression of tissue-restricted antigens within the thymus during negative selection,

self-reactive T cells still escape deletion. This is partially evidenced by occurrences of T cell mediated autoimmune disorders in individuals with otherwise healthy immune systems (Dornmair,

Goebels, Weltzien, Wekerle, & Hohlfeld, 2003). Within the subfield of antigen discovery for cancer immunotherapy, numerous T cell clonotypes have been isolated which recognize self-antigen expressed by tumours (Wurz, Kao, & DeGregorio, 2016).

1.3.2 Intracellular antigen processing

Antigen specificity of the adaptive immune response relies on the ability of host cells to present a given epitope in the context of MHC. Two primary antigen processing pathways dictate what epitopes may bind to MHCI and MHCII for presentation to CD8 T cells and Th cells, respectively.

(19)

9 Endogenous antigens are presented by MHCI to CD8 T cells

The primary function of MHCI is the presentation of endogenous antigen to CD8 T cells (Figure 1). This molecule is expressed on the surface of all nucleated cells of the human body and is crucial for consistent immune surveillance throughout the body. It consists of an α-heavy chain, ß2-microglobulin, and displays an endogenously-derived short peptide fragment within the extracellular groove (A. C. Goldberg & Rizzo, 2015). Within a given cell, cytosolic proteins, regardless of origin, are routinely tagged with ubiquitin, labeling the protein for proteasomal degradation (Dolan et al., 2012). Peptide fragments are subsequently transported to the endoplasmic reticulum (ER) by TAP (transporter associated with antigen processing), where they are further degraded by endoplasmic reticulum-associated aminopeptidases 1 and 2 (ERAP1 and 2) into short 8-11mer peptides ideal for binding MHCI (Kanaseki, Blanchard,

Hammer, Gonzalez, & Shastri, 2006; Serwold, Gonzalez, Kim, Jacob, & Shastri, 2002). Within the ER, chaperone proteins confer stability to the nascent MHCI and facilitate proper entry of the peptide into the binding groove. Once bound, the stabilized MHCI:peptide complex is

transported through the Golgi to the cell surface. This process allows CD8 T cells to monitor intracellular protein expression, in order to identify and delete cells harbouring proteins of a foreign or mutated nature (A. C. Goldberg & Rizzo, 2015). In the context of this thesis, this process is notable as it facilitates the presentation of tumour associated antigens (TAA).

(20)

10 Figure 1. Antigen presentation by MHCI to CD8 T cells. Intracellular proteins are digested by the proteasome and transported into the ER for MHCI loading. Loaded MHCI are shuttled to the cell surface for presentation to TCR of the appropriate specificity.

Exogenous antigens are presented by MHCII to CD4 T cells

The purpose of MHCII is to present of exogenous antigen to CD4 T cells. Unlike MHCI, MHCII is only expressed on professional APCs, such as B cells, dendritic cells (DCs), and

macrophages. Exogenous matter is taken up by APCs via endocytosis. Depending on the type of APC, endocytosis may be non-specific, as in innate immune cells such as DCs and macrophages. For B cells, this is triggered exclusively by antigen specific engagement of the B cell receptor. Once endocytosis of the exogenous material has occurred, the endosome fuses with a

(21)

11 lysosome, where the contents are digested into individual peptide fragments prior to being loaded onto MHCII. Due to the unstable structure of MHCII, when it is synthesized in the ER, it is bound to the ‘invariant chain’, CD74, which offers stability (A. C. Goldberg & Rizzo, 2015). The complex then leaves the ER in the MHCII compartment (MIIC). Within the MIIC, proteases cleave CD74, leaving behind the class II-associated invariant chain peptide (CLIP), which occupies the binding groove, maintaining its structure. Once MIIC and the peptide-containing endosome fuse, the chaperone protein human leukocyte antigen DM (HLA-DM) facilitates the replacement of CLIP by a foreign peptide, creating the MHCII:peptide complex (A. C. Goldberg & Rizzo, 2015). The complex can then be transported to the APC surface, where peptides of an exogenous origin are displayed to antigen specific CD4 T cells.

Together, MHCI and MHCII presentation of antigenic peptides can lead to the induction of an antigen specific immune response. These responses are directed towards any cell

expressing the antigenic protein.

1.3.3 Additional cells involved in the T cell immune response The CD8 immune response is supported by APCs and CD4 T cells

The cytotoxic activity of CD8 T cells is supported by a series of complex interactions with a variety of other immune cell types. Notably, APCs and CD4 T cells provide stimuli for initiation and maintenance of the cell mediated immune response. Initiation is frequently conducted by DCs, which ingest and degrade cellular debris in order to present antigen for the activation of

(22)

12 both CD8 and CD4 T cells. As mentioned, MHCI is classically reserved for the presentation of endogenous protein; however, cross-presentation pathways allow DCs to display exogenous antigen in order to instigate CD8 T cell activation. Cross-presentation by DCs allows the recruitment of peripheral antigen specific CD8 T cells which have not directly encountered a target cell population. To aid in the activation and maintenance of the CD8 T cell response, DCs also activate CD4 T cells. As professional APCs, DCs express and display MHCII, which facilitates antigen presentation to the CD4 T cell TCR, an interaction which is stabilized by CD4 (Knosp & Johnston, 2012). Activated CD4 T cells may differentiate into a variety of effector phenotypes, including Th1, Th2, Th17, and Treg cells, each of which possess unique functions. Of these, Th1 cells are the most directly related to CD8 directed immunity against intracellular pathogens. The activation of CD8 T cells is promoted by interleukin-2 (IL-2) and IFNγ, which are both secreted by Th1 cells. The differentiation to Th1-type immune responses is favoured in the presence of IL-12, IL-18, and IFNγ (Knosp & Johnston, 2012). Together, APCs and Th1 cells stimulate the immune system to mount a CD8 T cell mediated immune response.

The Th2 lineage of CD4 T cells promotes humoural immunity

The humoural immune response, mediated by B cells, is promoted by CD4 T cell differentiation to a Th2 phenotype. Commitment of CD4 T cells to the Th2 lineage is primarily due to the influence of IL-4. Subsequent IL-4, IL-5, and IL-6 production by Th2 cells serves to function in a positive feedback loop as well as promote B cell activation and B cell

differentiation into immunoglobulin-secreting plasma cells (Yang & Ansell, 2012). Plasma cells produce and release antibodies which facilitate antigen-targeted cell killing and inhibition via

(23)

13 phagocytosis, complement activation, or antibody-dependent cellular cytotoxicity (ADCC). Despite this, in anti-tumour immunity, the role of B cells has been controversial. For example, some studies have reported that B cells have immunosuppressive effects by directly inhibiting CD8 T cell responses (Olkhanud et al., 2011; Z. Qin et al., 1998), while others have reported that B and T cell tumour infiltrates can co-localize at tertiary lymphoid structures in order to mount cooperative anti-tumour immune responses that correlate with improved survival (Kroeger, Milne, & Nelson, 2016). Indeed, infiltration of plasma cells has reportedly been associated with favourable prognoses in a number of cancer types (Iglesia et al., 2014; Lohr et al., 2013;

Richards et al., 2012; Schmidt et al., 2012). Apart from their classical role in antibody

production, B cells also release immunomodulatory cytokines that promote T cell responses, possibly explaining some of the benefits of B and T co-infiltration. B cells also act as APCs, which can supplement DC activity to activate and support a CD8 T cell response. Together, CD4 T cells and B cells provide an additional avenue for the immune system to attack cancerous cells.

An alternate lineage for CD4 cells – Th17

At times in conflict with the other T subsets, CD4 cells may also differentiate into Th17 cells. Differentiation to the Th17 lineage is promoted by numerous cytokines including 6, IL-21, IL-23, and TGF-ß (H. Qin et al., 2009; X. O. Yang et al., 2007). Committed Th17 cells are partially defined by their ability to secrete IL-17, but subsets within the Th17 classification have variable effects on tumour immunity. As further discussed in Bailey et al (2014), these cells may complement or inhibit an immune response against cancerous cells. In the presence of IL-12, Th17 cells may further differentiate into Th1-like cells to strengthen a CD8 T cell-mediated

(24)

anti-14 tumour response (Wang et al., 2014). This alteration is characterized by a loss of IL-17

production and an initiation of IFNγ secretion. However, in the presence of TGF-ß, Th17 cells may assume a regulatory role, in which they upregulate FoxP3 expression, and also foster the recruitment of Treg cells (Bailey et al., 2014). As such, Th17 cells contribute to the modulation of the cell mediated immune response, playing a variable role between Th1 and Tregs.

Restraining the immune response – T regulatory cells

T regulatory cells are an immunosuppressive class of cells whose classical role involves mediating peripheral tolerance to self-antigens, which is critical to the prevention of

autoimmunity and resolving acute inflammatory reactions. Typically defined as

CD4+CD25+FoxP3+ T cells, Tregs inhibit the actions of other pro-inflammatory immune responses through mechanisms such as contact inhibition or secretion of immunosuppressive cytokines such as IL-10 and TGF-ß (Ulges, Schmitt, Becker, & Bopp, 2016). Like CD8 or CD4 cells, Tregs typically mature in the thymus; however, regulatory differentiation from naïve CD4 T cells may occur as well. This spontaneous Treg differentiation is induced by the cytokines IL-6 and TGF-ß (Kiraz, Baran, & Nalbant, 2016). While Tregs are critical to the prevention of devastating acute inflammation (e.g. cytokine storm) and autoimmunity, they also frequently infiltrate tumours and antagonize beneficial anti-tumour immune responses. As such, some studies have reported that the presence of intratumoural Tregs correlates with poor prognosis and

suppression of tumour immunity (Antony & Restifo, 2005; Bos, 2016). Curiously, other reports have found that Treg infiltration is positively associated with improved overall survival

(25)

15 have shown that Treg infiltration can be a sign of an active anti-tumour immune response – in ovarian cancer, tumours with high ratios of CD8 T cells to Tregs demonstrate superior overall increased overall survival (Sato et al., 2005). Unfortunately, the activity of Treg cells, which is critical during most immune responses, can be quite detrimental for the anti-tumour response.

Overall, the immune system involves a host of cell types which fill specific support and effector roles in order to provide an effective immune response against antigen of tumoural or foreign origin.

1.4 Cancer and the Immune System

1.4.1 Immunoediting – The evolution of a tumour

If the immune system is capable of recognizing and eliminating cancerous cells, why are cancerous cells still able to grow and establish tumours, invade and metastasize? In 2004, Dunn and colleagues proposed their model of cancer immunoediting, which revolves around three concepts: elimination, equilibrium, and escape (Dunn, Old, & Schreiber, 2004). Each concept of this model contributes to a stage of malignant tumour genesis and its interaction with

regulatory immune oppression.

Elimination – The immune system kills newly mutated cells

The first phase of this model, elimination, describes the initial generation of a mutating cellular mass. It is triggered by the initial mutations which have led a given cell to circumvent the genetic reins which maintain regular proliferation patterns (Macleod, 2000). Accelerated

(26)

16 proliferation of the mutated cells cause physical tissue disruption of the local environment (Carmeliet & Jain, 2000; Hanahan & Folkman, 1996; Sternlicht & Werb, 2001). This disruption triggers an inflammatory response which recruits an innate immune response against the new tumour (Vicari & Caux, 2002). Through the actions of the innate immune system, macrophages and NK cells cause IFNγ production and subsequent tumour killing (Bromberg, Horvath, Wen, Schreiber, & Darnell, 1996; Coughlin et al., 1998; Kumar, Commane, Flickinger, Horvath, & Stark, 1997; Luster & Leder, 1993; Z. Qin & Blankenstein, 2000). Dead and dying tumour cells provide a source of tumour antigen: proteins to be ingested by DCs to prime T cell responses.

Activated DCs migrate toward the tumour draining lymph node after tumour antigen acquisition for presentation to naïve CD8 T cells (Gerosa et al., 2002). Chemokine gradients then attract newly activated T cells to the tumour site for killing. At the tumour site, cytokines provided by CD4 T cells help maintain the efficacy of tumour-specific CD8 T cells, induce tumour cell killing by mechanisms previously discussed. Dunn et al propose that this process repeatedly occurs throughout life (2004), as new adaptive responses are directed toward each

antigenically distinct tumour. When the tumour has been successfully eradicated by the

immune response, this process ends. However, if the tumour is not killed, it proceeds to a state of equilibrium.

Equilibrium – A balance between tumour proliferation and immunogenic death

The equilibrium phase of tumourigenesis may persist indefinitely and is resolved by either eradication or escape. It is a state of homeostasis in which the immune system places an adaptive evolutionary pressure upon the tumoural cells (Koebel et al., 2007). Throughout,

(27)

17 genetically unstable cells continually proliferate into antigenically distinct generations of

potentially malignant progenitors (Lengauer, Kinzler, & Vogelstein, 1998). As in macroevolution, cells that are unfit are unable to survive the attack of the immune system, while those that do survive produce a new generation of daughter cells carrying the mutations of their parental cell. Growing tumours fluctuate between times of increased proliferative rates, when a beneficial mutation occurs, and increased death rates, when the immune system identifies novel tumour antigens. It is only after random mutations accumulate over generations that a tumour cell might have the capability to evade the adaptive immune system – this may take many years (Loeb, Loeb, & Anderson, 2003). When the immune system can no longer identify nor kill tumour cells at a competitive rate, the tumour has escaped the balance between growth and immune directed death, allowing it to proliferate in spite of immunological pressures (Khong & Restifo, 2002).

Escape – The tumour overcomes immune pressure

When the tumour has escaped the immune response, the tumour may proliferate to the point of being clinically detectable (Schreiber, Old, & Smyth, 2011). In the escape stage, a tumour may have acquired various abnormalities that contribute to immune evasion. These may include any of, but are not limited to: loss and down-regulation of MHCI, defective death receptor signaling, cytokine-directed immunosuppression, FasL directed T cell apoptosis. Indeed studies have shown that over 40% of tumours downregulate MHCI, which avoids T cell immune recognition and destruction (Algarra, Collado, & Garrido, 1997; Marincola, Jaffee, Hicklin, & Ferrone, 2000). Additionally, many tumours display defects in the antigen processing

(28)

18 pathway for MHC presentation (Korkolopoulou, Kaklamanis, Pezzella, Harris, & Gatter, 1996; Restifo et al., 1993; Seliger et al., 1997). These alterations inhibit recognition or killing by the immune system, allowing tumour cells to continue proliferating.

1.4.2 Immunotherapy – Helping the immune system fight cancer

The purpose of immune-based therapies is to aid the natural immune system in the conflict with a continually evolving tumoural population. A large portion of the cancer immunotherapy field focuses on improving the presence and activity of T cells in the tumour environment. While tumour infiltrating lymphocytes (TIL) have been identified as early as the 1860s (Virchow, 1863), our understanding of their role in tumour control has only recently appreciated. In many cancer types, including ovarian, T cell infiltration of the tumour is

correlated with positive clinical outcomes (Hamanishi et al., 2007; Mariya et al., 2014; Milne et al., 2009; Sato et al., 2005; Zhang et al., 2003). Although this correlation does not mean that TIL directly mediate clinical benefit, it has been found that many of these infiltrative T cells are indeed tumour specific (Lu et al., 2014). Since this discovery, immunotherapy has become a frequent subject of research. Due to the complex nature of cancer immunology,

immunotherapeutic strategies are broad and aim to influence the activity of the immune

system in diverse ways. The strategies discussed here were selected based on their relevance to this thesis and their prevalence within the field. Each of these strategies has shown some therapeutic promise: vaccines, adoptive cell therapy (ACT), chimeric antigen receptor (CAR) T cells, as well as immune checkpoint blockades.

(29)

19 Vaccines – Providing tumour antigens for immune activation

Therapeutic cancer vaccines aim to direct immune responses by providing tumour antigens to the immune system in order to establish a tumour-specific memory population. Tumour antigens can be provided in a number of different ways: peptides or complete proteins, DCs, recombinant viral vectors, and tumour cells, amongst others (Schlom, 2012). Peptide vaccines typically consist of a set of peptides representing a small number of tumour antigens, and are supplemented with adjuvant on various prime-boost schedules. Dendritic cell vaccines provide a vehicle for the delivery of selected epitopes to the patient, as they are potent antigen presenting cells capable of generating a robust immune response directed towards the tumour antigen. Similarly, viral vector vaccines are tailored to specifically encode the antigen of interest prior to infection of the host. Tumour antigens are then expressed within antigen presenting cells for stimulation of the host immune response. Selection of the target tumour antigen is one of the greatest challenges within the development of any of these strategies. For a given target to give rise to a viable tumour epitope, intracellular processing machinery must naturally process a unique the antigenic epitope from the parent protein, the tumour-specific peptide must have the threshold affinity required to bind MHC, and T cells must exist within the immune repertoire which have sufficient affinity and specificity for the antigen (Martin et al., 2016). In addition, an ideal target would confer a survival advantage to the tumour cells expressing it, thus preventing target-downregulation without consequence.

Identification of tumour antigens for immunological targeting poses a difficult obstacle for vaccination strategies. However, autologous tumour cell vaccines bypass the antigen

(30)

20 selection problem. This strategy has the capacity to apply broad spectrum activation against any tumour antigens present. Autologous tumour cell vaccines are often supplemented with transduced granulocyte-macrophage colony-stimulating factor (GM-CSF) to enhance APC functionality (D. Z. Chang, Lomazow, Joy Somberg, Stan, & Perales, 2004; Simons & Sacks, 2006; Yu, Chueh, Tsai, Lin, & Qiu, 2016), or even fused with autologous DCs directly, in order to

bypass the process of initial uptake of the antigen (Garcia-Marquez, Shimabukuro-Vornhagen, Theurich, & von Bergwelt-Baildon, 2013). However, T cells exhibit competitive binding to DCs for activation, which may prevent diverse T cell activation to such a variety of antigen, whereas DCs presenting a single tumour antigen to a single T cell clone may be more efficacious.

Adoptive cell therapy – Providing immune cells for tumour attack

While vaccination aims to target a tumour based on a single or small group of specific antigens, ACT is geared toward broadly targeting any antigens that may be present on the tumour. With this method, TIL are harvested from the tumour and cultured ex vivo under conditions which may somewhat vary between studies. For example, they may be stimulated by autologous DCs or allogeneic irradiated mononuclear cells, in conjunction with anti-CD3 antibody and IL-2. In this manner, TIL cultures expand to the order of 1011 cells, and are then infused back into the patient after lymphodepletion, which enhances the effector function of infused antigen specific T cells (Gattinoni et al., 2005; Rosenberg & Restifo, 2015). This approach has yielded significant clinical responses by tumours, particularly in melanoma patients (Besser et al., 2013; Dudley et al., 2002; Pilon-Thomas et al., 2012; Radvanyi et al.,

(31)

21 2012). There are now multiple ovarian cancer clinical trials assessing the merits of this

technique (NCT01883297, NCT01174121, NCT02482090).

Engineered T cell receptors – Creating antigen specific T cells

Engineered TCR therapy shares many aspects of each of the strategies discussed above. Like vaccination, cloned TCRs focus on single or few known antigens. These antigens must be known in order to generate an antigen specific TCR. A bulk population of antigen specific T cells can then be infused to the patient in a manner identical to ACT. In this technique, TCR

specificity is selected and genetically directed for the infused T cell product. These genetic alterations can be made virally with high efficiency (Rosenberg & Restifo, 2015). This can grant the T cell repertoire the ability to recognize antigens which there is existing peripheral

tolerance to or to which TCR frequency is undetectably low. In another technique, using chimeric antigen receptor (CAR) T cells, immunoglobulin variable regions may be linked to the intracellular domain of the TCR. This directs T cell cytotoxicity against antigens which are not limited to MHCI:peptide complexes. These new antigens may simply be irrelevant surface proteins expressed uniquely in tumoural tissue (Imai et al., 2004; Song et al., 2011). For

example, CD19 CAR T cell therapy has shown remarkable successes in B cell lymphomas (Maude et al., 2014; Porter, Levine, Kalos, Bagg, & June, 2011).

Checkpoint blockade – Taking off the brakes

Debatably the most popular of current immunotherapy tactics, immune checkpoint blockade has shown promising results in various cancer types, including melanoma NSCLC, and

(32)

22 renal cell carcinoma. Within this class of therapy, the cytotoxic T-lymphocyte-associated protein 4 (CTLA-4) inhibitor ipilimumab and PD-1 inhibitors nivolumab and pembrolizumab have been recently approved by the FDA. As discussed in section 1.3, CD28 binding by the APC molecule B7 is required for CD8 T cell activation. However, T cell activation also causes upregulation of CTLA-4, which intercepts B7 binding to CD28. Engagement of CTLA-4 blocks costimulation and inhibits T cell activation (Chambers, Kuhns, Egen, & Allison, 2001; Teft, Kirchhof, & Madrenas, 2006). Ipilimumab blocks this inhibitory interaction, allowing T cell activation to proceed. Indeed, ipilimumab has demonstrated particular efficacy in the context of melanoma, with rates of response ranging up to 25% (Hodi et al., 2010; Schadendorf et al., 2015). Additionally, studies have revealed significant clinical benefit in a number of other cancer sites, including non-small cell lung cancer, renal cell carcinoma, prostate cancer, and ovarian cancer (Hodi et al., 2008; Lynch et al., 2012; van den Eertwegh et al., 2012; J. C. Yang et al., 2007). Similarly, the receptor programmed cell death-1 (PD-1) and its ligand PD-L1 have also been the target of numerous antibody based blockade strategies. Expression of PD-L1 can be stimulated by IFN-γ in many cell types, including tumour cells, which in turn engage PD-1 receptors on activated T cells. Binding of PD-1 leads to attenuation of the T cell response. Anti-PD-1/PD-L1 therapies have demonstrated significant clinical benefit in many different cancerous settings (Ansell et al., 2015; Brahmer et al., 2015; Gettinger et al., 2015; Le et al., 2015; Motzer et al., 2015), with ongoing trials (NCT02626065, NCT02504372, NCT02853331, NCT02819518, NCT02684292, NCT02702401, and more). Although both checkpoint blockade strategies have shown

significant clinical benefit, neither have substantial claims to the mechanism of action behind their effects. However, they have common prognostic predictors, which include a correlation

(33)

23 with mutational load, high TIL frequency, as well as the presence of their respective targets on TIL.

1.5 Lactate Dehydrogenase C as a Tumour Antigen 1.5.1 Cancer Testis Antigens

Despite the efficiency of thymic selection at regulating self-tolerance within the immune system, it is not infeasible for T cells to escape these mechanisms (Dornmair et al., 2003). It is known that autoreactive T cells are present in the periphery. These are the autoreactive T cells we wish to isolate in order to characterize a naturally expressed protein as a TAA. When choosing potential antigens to investigate, tumour specificity is of great importance. For this reason, the upregulation of testis-restricted proteins within cancerous cells is notable. These so called cancer-testis (CT) antigens, of which there are over 200, constitute a significant portion of all identified potential TAA. Many of these CT antigens have been investigated by other groups (Vigneron, Stroobant, Van den Eynde, & van der Bruggen, 2013), and reactive epitopes have been categorized and defined by the MHC allele they were discovered on.

Of these CT antigens, NY-ESO-1 and MAGE-A3 have made a notable impact (Wurz et al., 2016). Studies of NY-ESO-1 targeted immunotherapies, primarily conducted in melanoma patients, have shown significant efficacy, with positive response rates as high as 88%. These clinical trials vary in method, but include adoptive CD4 T cell therapy, adoptive CD8 T cell therapy, NY-ESO-1-expressing viral vaccination, and peptide/adjuvant vaccination.

(34)

24 Immunotherapies based on MAGE-A3 have not shown such promising results. While protein and peptide vaccinations have had promise in early phase clinical trials, no phase III studies have upheld their efficacy. Furthermore, a phase I/II trial of an adoptive cell therapy with engineered TCRs against MAGE-A3, despite generating long-lasting responses in multiple patients, exhibited cross-reactivity with MAGE-A12 (Morgan et al., 2013). This cross reactivity led to substantial neuronal destruction and the death of two patients. As such, it is critical that additional targets continue to be evaluated as potential antigens.

The testis-restricted LDH isoform lactate dehydrogenase C (LDHC), provides an intriguing potential immunotherapeutic target, given its expression patterns. Apart from its testis

restriction, one study has observed its ectopic expression in melanoma, breast, colon, lung, ovarian, prostate, renal, thyroid, and cervical cancers (Koslowski et al., 2002). Broad expression between patients across multiple cancer types makes LDHC a tempting prospect for antigen discovery.

1.5.2 Lactate dehydrogenase

Enzymatic activity of lactate dehydrogenase

Lactate dehydrogenase (LDH) is a tetrameric enzyme which catalyzes the conversion between pyruvate and lactate, which is coupled with the conversion between NADH and NAD+. As the catabolic pathway of glycolysis utilizes the reduction of NAD+ to NADH during the

(35)

25 production. When sufficient oxygen is available, pyruvate enters the citric acid cycle in the mitochondrion, in order to supply NADH to oxidative phosphorylation. Oxidative

phosphorylation utilizes the oxidation of NADH to NAD+ in a series of coupled redox reactions to generate a proton gradient across the mitochondrial inner membrane which in turn is used in the generation of ATP. However, when oxygen is not available to the cell, ATP is solely generated by glycolysis. The reduction of pyruvate by lactate dehydrogenase provides

regeneration of NAD+ to be utilized in further glycolytic reactions for additional ATP generation. Utilizing this pathway, a cell can generate ATP at a significantly higher rate than those that rely on oxidative phosphorylation (Pfeiffer, Schuster, & Bonhoeffer, 2001). However it cannot be maintained for long, as it increases the local concentration of lactic acid, as commonly noted in muscles during anaerobic exercise. In the liver, lactate dehydrogenase converts lactate back to pyruvate, which is recycled in gluconeogenesis, to be returned to muscle and other cells as glucose (Kalderon, Korman, Gutman, & Lapidot, 1989). Together, the pathways of anaerobic glycolysis and gluconeogenesis form the Cori cycle.

The isoenzymes of lactate dehydrogenase

The monomer subunits which together assemble the LDH tetramer are encoded by the three genes LDHA, LDHB, and LDHC. Together, subunits A and B may generate any of five unique homo- and hetero- tetramers, which are differentially assembled in various tissues, although general expression of the LDHA and LDHB genes is fairly ubiquitous throughout the body (Blanco & Zinkham, 1963). The LDH gene LDHC encodes a third isoform, which is involved in the assembly of the homotetramer LDH-C4 (E. Goldberg, Eddy, Duan, & Odet, 2010). For the

(36)

26 purposes of this thesis, it is this gene product which will be here on referred to as LDHC. The expression of LDHC is restricted to male germ cells (E. Goldberg et al., 2010; Koslowski et al., 2002). LDHC, having developed due to duplication of the LDHA gene, is structurally extremely similar to both LDHA and B. Where its differences lie, however, are in enzymatic activity, due to select differences in its primary sequence (Figure 2). Amino acid substitutions, strewn

throughout the protein and within the active site, provide alterations which to contribute to its unique activity within the testis. These alterations lead to a difference in substrate/product inhibition patterns from LDHA and B (Li, Fitch, Pan, & Sharief, 1983; Li et al., 1989). Each of these isoforms experiences some level of substrate inhibition from pyruvate, LDHC being the most sensitive (Battellino, Jaime, & Blanco, 1968). While catalyzing the reverse reaction, lactate to pyruvate, which is thermodynamically less favourable, LDHC is not inhibited by increased concentrations of lactate, unlike LDHA. However, these studies observed this inhibitory activity at concentrations reaching far beyond that of what would be seen in the body (Battellino et al., 1968; Usher-Smith, Fraser, Bailey, Griffin, & Huang, 2006). This pattern of inhibition has, in the past, led investigators to believe that LDHC preferentially catalyzes the reverse reaction of lactate oxidation. This is, in fact, not the case, as numerous enzymatic studies have shown that each LDH catalyzes this reversible reaction without bias, albeit with different rates (Battellino et al., 1968). A number of studies have also shown that the differential amino acids of LDHC grant it broader substrate specificity – LDHC may additionally catalyze the redox reactions between 2-oxo-butanoate/2-oxo-pentanoate and 2-hydroxy-butanoate/2-hydroxy-pentanoate, whereas the ubiquitous enzyme isoforms cannot (Blanco, 1973). What metabolic advantages this

(37)

27 flexibility may provide spermatozoa in the testis are currently unknown, although LDHC activity is crucial to the development of spermatozoa (Odet et al., 2008).

LDHC MSTVKEQLIEKLIEDDENS-QCKITIVGTGAVGMACAISILLKDLADELALVDVALDKLK LDHA MATLKDQLIYNLLKEEQTP-QNKITVVGVGAVGMACAISILMKDLADELALVDVIEDKLK LDHB MATLKEKLIAPVAEEEATVPNNKITVVGVGQVGMACAISILGKSLADELALVDVLEDKLK LDHC GEMMDLQHGSLFFSTSKITSGKDYSVSANSRIVIVTAGARQQEGETRLALVQRNVAIMKS LDHA GEMMDLQHGSLFLRTPKIVSGKDYNVTANSKLVIITAGARQQEGESRLNLVQRNVNIFKF LDHB GEMMDLQHGSLFLQTPKIVADKDYSVTANSKIVVVTAGVRQQEGESRLNLVQRNVNVFKF LDHC IIPAIVHYSPDCKILVVSNPVDILTYIVWKISGLPVTRVIGSGCNLDSARFRYLIGEKLG LDHA IIPNVVKYSPNCKLLIVSNPVDILTYVAWKISGFPKNRVIGSGCNLDSARFRYLMGERLG LDHB IIPQIVKYSPDCIIIVVSNPVDILTYVTWKLSGLPKHRVIGSGCNLDSARFRYLMAEKLG LDHC VHPTSCHGWIIGEHGDSSVPLWSGVNVAGVALKTLDPKLGTDSDKEHWKNIHKQVIQSAY LDHA VHPLSCHGWVLGEHGDSSVPVWSGMNVAGVSLKTLHPDLGTDKDKEQWKEVHKQVVESAY LDHB IHPSSCHGWILGEHGDSSVAVWSGVNVAGVSLQELNPEMGTDNDSENWKEVHKMVVESAY LDHC EIIKLKGYTSWAIGLSVMDLVGSILKNLRRVHPVSTMVKGLYGIKEELFLSIPCVLGRNG LDHA EVIKLKGYTSWAIGLSVADLAESIMKNLRRVHPVSTMIKGLYGIKDDVFLSVPCILGQNG LDHB EVIKLKGYTNWAIGLSVADLIESMLKNLSRIHPVSTMVKGMYGIENEVFLSLPCILNARG LDHC VSDVVKINLNSEEEALFKKSAETLWNIQKDLIF- LDHA ISDLVKVTLTSEEEARLKKSADTLWGIQKELQF- LDHB LTSVINQKLKDDEVAQLKKSADTLWDIQKDLKDL

Figure 2. Alignment of LDHC to LDHA and LDHB. Amino acid sequences of LDHC, LDHA, and LDHB. Amino acid residues which are unique to LDHC (62/332) are highlighted.

Lactate dehydrogenase activity in the tumour

In addition to in the testes, LDHC is frequently upregulated in cancer of various types. Within the tumour, while LDHC may specifically confer benefits unique to its particular isoform, general LDH expression does aid in tumour growth and survival. LDH expression allows

cancerous cells to generate ATP via anaerobic glycolysis, which is advantageous for a number of reasons. A defining characteristic of cancerous cells is their tendency to proliferate at a

(38)

28 substantially increased rate, which requires a significant amount of biomaterial. That material is provided by mid-cycle substrates in the tricarboxylic acid (TCA) cycle (Vander Heiden, Cantley, & Thompson, 2009). Due to the loss of TCA cycle substrates, oxidative phosphorylation is largely unrealistic; thus, tumours frequently generate ATP by anaerobic respiration. Furthermore, the binary nature of oxygen availability within the tumour lends itself to the requirements of LDH, as hypoxic regions of the tumour can convert pyruvate to lactate, which can be taken up by more oxygenated regions of the tumour for oxidative phosphorylation (Sonveaux et al., 2008). Additionally, the lactic acidosis of the tumour microenvironment is inherently anti-immunogenic, as it compromises the functionality of T cells (Blank, Haanen, Ribas, & Schumacher, 2016). This suppression of function is also evidenced by the strong inverse correlation of serum LDH concentration with the clinical outcomes of PD-1 and CTLA-4 immune checkpoint blockades (Diem et al., 2016). Due to these characteristics which are beneficial to the tumour, LDHC makes a tempting potential antigen: in addition to the immune response being targeted against the tumour, it would be targeted specifically due to a protein which provides a degree of cellular fitness to the tumoural clones expressing it – a quality typically attributed to driver mutations.

1.6 Summary

This chapter has illustrated the interaction between the immune system and cancer in general. The highly evolved adaptive immune system subjects a developing tumour to

(39)

29 escape occurs when the immune system has failed to contain tumour development and tumour evolution has developed the ability to grow at rate which has surpassed the rate of immune elimination. The activity of CD8 T cells in immunoediting and tumour control is critical and is supported by and shared with a diverse set of immune cells, of an innate origin as well as an adaptive origin. Immunotherapy posits to bolster the immune response to cancer by acting against some common immune evasion methods, such as antigen escape. In this past chapter I have described some of the methods used with this aim. One recurring challenge that some immunotherapeutic strategies run into is the selection of immunological targets within a given tumour. Here I propose lactate dehydrogenase C as an immunological CD8 T cell target. My hypothesis was that T cells exist within a patient’s T cell repertoire which can recognize LDHC peptides in the context of MHCI, and kill autologous cells expressing LDHC.

(40)

30

Chapter 2: LDHC peptide specific T cells reside in the patient T cell

repertoire

David S. Neilson1,2, Darin A. Wick1, Jennifer L. Kalina1, Spencer D. Martin1,3,4, David Kroeger1, Julie S. Nielsen1,2, John Webb1,2, Zoe Petropoulos1,2, Luke Hughson1,2, Julian J. Lum1,2

1

Trev and Joyce Deeley Research Centre, BC Cancer Agency, Victoria, BC, Canada 2

Department of Biochemistry and Microbiology, University of Victoria, Victoria, BC, Canada 3

Michael Smith's Genome Sciences Centre, Vancouver, BC, Canada 4

Interdisciplinary Oncology Program, University of British Columbia, Vancouver, Canada

DSN, DAW, and JJL designed the study. DSN, DAW, JLK, SDM, DK, JSN, JW, and JJL were involved in data acquisition and/or interpretation. ZP and LH provided preliminary data.

(41)

31

2.1 Introduction

Interaction between the immune system and cancer is dynamic and complex (Dunn et al., 2004). Throughout life, tumour control is mediated by this interaction. Regardless of the surrounding complexities, studies have shown that CD8 T cell infiltration is associated with improved outcome in most solid cancer types (Fridman et al., 2012; Milne et al., 2012). Recent studies in immune blockade therapy suggest that lymphocytic infiltrates recognize tumour antigens, but are often inhibited through various means within the tumour microenvironment (Blank et al., 2016; Weber, 2010). Despite the knowledge of T cell tumour infiltration and its prognostic relevance, TCR specificities of T cell infiltrates are relatively unknown. However, studies have revealed that T cells specific to known antigens do account for a small portion of all CD8+ infiltrates, and there are indeed numerous accounts of TIL which exhibit tumour specificity. These accounts have shown that subsets of CD8+ TIL can both recognize and kill autologous tumour cells in vitro (Andersen et al., 2012; Kelderman et al., 2016).

In the field of cancer immunotherapy, there is a constant need to discover novel antigens. To mount an anti-tumour immune response, CD8 T cells require tumour antigens, derived from mutated or tissue specific proteins. However, on the spectrum of immunogenic cancer types, HGSC ranks relatively low, with a low mutational burden (Alexandrov et al., 2013). In melanoma, proven to be a highly immunogenic cancer, a portion of aggregate TIL are specific to somatic point mutations arisen throughout tumour evolution. In contrast, recent studies of HGSC have revealed that very few cases (12%) are likely to provide point mutation-derived immunogenic neoantigens for therapy (Martin et al., 2016). For these reasons, it is imperative

(42)

32 that projects are undertaken to identify additional tumour targets for immunotherapy. Cancer-testis proteins have shown viability as CD8 tumour targets, as demonstrated in therapeutic studies involving MAGE proteins and NY-ESO-1 (Wurz et al., 2016). These CT antigens are particularly intriguing targets in the context of ovarian cancer, as they are tumour specific and also happen to be expressed in a high proportion of tumours. Antigenic tumour proteins expressed across multiple patients present an opportunity to generate a more widely

applicable potential therapy. Notably, the CT antigen LDHC has been found to be expressed in patient tumours at frequencies as high as 100% (lung adenocarcinoma, n=18), through 83% (cervical cancer, n=6), 44% (melanoma, n=16), to as low as 35% (breast cancer n=20) (Koslowski et al., 2002; Yen et al., 2007). In our cohort of patients, preliminary PCR data revealed that 76% of HGSC patients express tumoural LDHC (22/29, data not shown). Such high frequency of a potential antigen warranted further investigation. As such, I hypothesized that HGSC patients harbour CD8 T cells which recognize individual MHCI epitopes unique to LDHC.

To investigate this hypothesis, I prioritized the analysis of HGSC patients whose tumours express LDHC mRNA. Ascites from LDHC+ ovarian cancer patients was utilized for analysis due to the origin of ascites fluid. Ascites is a peritoneal build up of fluid very common in ovarian cancer patients, which contains slough from the tumour. Tumour associated lymphocytes (TAL) from the tumour slough may contain antigen specific T cells which have previously infiltrated the tumour.

When mining the T cell repertoire with the aim of identifying a specific T cell clone, it is crucial to consider its availability. While a given activated clone may occupy a substantial

(43)

33 portion of the lymphocytic compartment, naïve antigen specific T cells may be present at

frequencies no higher than 1/105. Considering this potential infrequency of a LDHC specific T cell, it had to be ensured that rare T cell populations are detectable. Bulk T cell expansion methods from tumour infiltrate can be used for the assessment of both TIL and TAL; however, these methods provide no degree of enrichment for rare antigen specific T cells. These methods sustain a given T cell clone with a single origin cell at a frequency of 1/105, and the threshold of detection of the methods at our disposal was closer to 1/104. However, selective expansion methods may be and are frequently used in the literature for this reason. Expansion of a select T cell clone may be facilitated by DC stimulation, in which peptide-loaded autologous DCs repeatedly stimulate and expand specific T cells from a bulk T cell culture (Nielsen et al., 2016). However, due to the infrequency of DCs within peripheral blood, maturation of a suitable DC contingent requires autologous peripheral blood mononuclear cells (PBMC) in excess of 2x108 cells. In this instance, these enrichment methods proved to be infeasible for this project, as the accessibility of patient PBMCs was low in comparison.

Patient cells must be efficiently analyzed without using an inappropriate amount of material, as less is typically available from cancer patients. Recent studies have published one such viable method for identifying rare antigen specific T cells (Geiger, Duhen, Lanzavecchia, & Sallusto, 2009; Theaker et al., 2016). In this method, a bulk population of T cells is distributed amongst hundreds of individual cultures. These divided T cell cultures are simultaneously non-specifically stimulated: each culture originates from 2000 autologous T cells and is expanded in a parallel fashion. Provided that individual T cell clones grow at the same relative pace in each

(44)

34 culture, this method generates a series of cultures with a minimal T cell clone frequency of 1/2000. This frequency is well within the bounds of detection for immunological assay. Previous studies have demonstrated this method’s efficacy in the assessment of naïve T cell populations directed toward specific antigens such as gp100 in melanoma and cadherin-3 in breast cancer (Theaker et al., 2016). In each of these cases, tumour killing was successfully mediated by these isolated T cells. Given these results, this miniline strategy could be effective for the analysis of LDHC reactivity in HGSC cases.

In order to assess a given immunological target, it is common to evaluate a T cell population’s ability to recognize exogenous peptides potentially specific to their TCR. In these assays, synthetic peptides are provided in vitro for extracellular loading onto the MHCI. As such, inquisitor peptides must have a high affinity for the MHCI, which varies between patients in accordance with their human leukocyte antigen (HLA) haplotype – i.e. which set of MHCI alleles that patient carries. Binding prediction algorithms for MHCI such as NetMHCpan can help select specific peptides for assessment. This is an especially beneficial strategy in the assessment of patient-specific mutations. However, in the broad assessment of a shared CT antigen, it was prudent to utilize a method unbiased toward patients possessing any one specific individual HLA allele. With this objective, it was not possible to study just a select few peptides from the 1200+ possible epitopes of LDHC. Fortunately, T cells have been shown to recognize cognate epitopes contained within long 15mer peptides (Fiore-Gartland et al., 2016). Furthermore, a recent study has shown that T cell responsiveness is unaffected by the presence of irrelevant peptides within pools of up to 300 peptides (Chevalier et al., 2015). Thus, a LDHC library of

(45)

35 overlapping 15mer peptides was synthesized for assessment of T cell responsiveness in the context of peptide pools.

In this chapter, CD8 T cells from the ascites of five HGSC patients are screened for recognition of LDHC via a 15mer peptide library. A single CD8 T cell clone which recognizes LDHC peptide was isolated from one of five patients. This T cell clone was isolated by fluorescence-activated cell sorting (FACS) based on CD8 T cell activation marker 4-1BB for further analysis. Four of five patients did not show reactivity to the LDHC peptide library.

(46)

36

2.2 Methods

2.2.1 Patients, biospecimens, and clinical data

The ovarian cancer patient cohort from which biospecimens were drawn has been previously described in reports associated with our laboratory (Castellarin et al., 2013; Wick et al., 2014). In brief, collections occurred through a prospective study, operating in partnership with the BC Cancer Agency's Tumour Tissue Repository. The study was granted ethics approval by the Research Ethics Board of the BC Cancer Agency as well as the University of British

Columbia (Certificate REB# H07-00463). Participants provided informed written consent prior to collection of their samples and associated clinical data. Patients were diagnosed with HGSC and underwent the standard treatment consisting of surgery followed by carboplatin-based

chemotherapy with or without paclitaxel. To work with these human samples, I have taken the Tri-Council Policy Statement 2: Course on Research Ethics, which is mandatory for Tri-Council funded researchers participating in work with human samples.

Patient ascites samples were collected during primary surgery. Ascites cells were then isolated by centrifugation and cryopreserved in 50% fetal bovine serum (FBS), 40% complete media (Roswell Park Memorial Institute media [RPMI] 1640 [Thermo Fisher Scientific, Nepean, ON, CA] containing 10% FBS, 25 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid [HEPES], 2 mM L-glutamine, 50 μM β-mercaptoethanol, and 1 mM sodium pyruvate) and 10% dimethyl sulfoxide (DMSO; Sigma-Aldrich, St. Louis, MO). Ascites cell preparations were stored in liquid nitrogen vapour.

(47)

37 2.2.2 Peptide library generation

A peptide library of LDHC was designed based on the sequence described in the National Center for Biotechnology Information (NCBI) protein database (NP_059144.1). Known exon splice variants were accounted for as well, provided that mRNA expression had been verified by previous studies (Koslowski et al., 2002), and that variant exon splicing leads to the generation of novel amino acid sequences unaccounted for by the primary library (EAW68393.1,

AEW43814.1). The library was designed as a series of overlapping 15mer peptides (Table 1). Each peptide overlapped its adjacent peptides by 11 amino acids. Three additional overlapping 15mer peptides account for each of two unique junctions expressed in the variant sequences. Peptides were synthesized commercially (GenScript, Piscataway, NJ). Lyophilized peptides were reconstituted at 10 µg/mL in DMSO (Sigma-Aldrich).

A minimal peptide library corresponding to peptide 62 of the LDHC library was also designed (Table 2). This library consists of all 8-11mer peptides contained within LDHC peptide 62 as well as the complementary peptides of LDHA, sequence as described in the NCBI protein database (CAG46515.1). Peptides were commercially synthesized (GenScript) and reconstituted at 10 µg/mL in DMSO.